Peter Yang

@peteryang

How to Pick the Right Metric to Grow for Your Product

Picking the right metric to move for your product is hard. Well chosen metrics can motivate and align your team around a goal. Poorly chosen metrics can lead to bad trade-offs that hurt your product, company and customers.

Here are three questions to ask when picking a metric for your product:

If you enjoy reading this post, I’m writing a book for new and aspiring product managers to help them get a head start in their PM career. Sign up at principles.pm to get a free chapter now.

1. Is growing the metric good not just for your product but also for your company and customers?

It’s worth repeating what I wrote before:

Being metrics obsessed is not the same as being customer obsessed.

A common pitfall is when the goals of your product and your company are not aligned. For example, a feed team for a video site could set interactions (comments, likes, etc) as a metric to grow. This might lead to product decisions like always showing expanded comments in feed. But that could easily hurt the overall video site’s metric of minutes watched because the videos themselves will have less visibility.

An even worse pitfall is when the goals of your product and company are aligned, but growing the metric is bad for your customer. For example, suppose you were in charge of user profiles for a social network. You set a goal to grow profile completeness — that is, the percent of user profiles that have all their information filled out (e.g. education, job history). The easiest way to do that is to just constantly ask users questions like: “What job do you have?” “What school did you go to?” Your completeness metric skyrockets and you’ve given the social network more data for advertisers to target. But how do your users feel about constantly being prompted for their private info and having that info show up in ads without their consent?

2. Is the metric easy to understand and measure?

Can you explain this metric to your CEO in less than 30 seconds? If the metric starts declining, can you break it down to figure out why?

A common pitfall is selecting an overly complicated metric. For example, suppose you’re in charge of growing drivers for Uber. You have data showing that 20% of drivers (let’s call them full time drivers) are responsible for 80% of rides. You see that only 10% of part time drivers are converting to full time, so you decide to set a goal to double that to 20%. But is a 10% conversion rate good or bad? What happens if part time to full time conversion doubles but 90% of drivers who are converted stop driving altogether after a month? What happens if you convert riders to full time drivers, but those riders were never part time drivers? You could’ve saved yourself a lot of headache by setting a simple output goal for growing the number of full time drivers. You can always track conversion rates and other numbers as inputs, but you should try to limit yourself to one simple output metric.

3. Can your team directly grow this metric?

Is this metric directly under your team’s control? If you run an experiment, can you show statistically significant results quickly? Is this something that your team is excited about growing?

A common pitfall is selecting a metric that your team can’t move directly (talk about demoralizing your engineers!). For example, you might want to grow user views of your product as a goal. But if 90% of your product’s views are coming from feed and the feed team’s not aligned, you’re going to have a hard time. Can you find new discovery channels? Or can you convince the feed team that giving your product better discovery is also good for their numbers? This brings me to the final point:

Finding a great metric to grow is a shared process

You must get your team, your managers, and any adjacent teams to internalize why you selected this metric for it to be effective. Make sure that you have a clearly defined problem statement and vision for your product before trying to pick a metric to grow and walk through the above questions with your stakeholders to help validate whether the metric is a good fit.

Remember: Nothing’s more dangerous than successfully growing the wrong thing.

I’m writing a book for new and aspiring product managers. If you enjoyed this post, visit principles.pm to get a free chapter.

More by Peter Yang

Topics of interest

More Related Stories